Message boards : Number crunching : Hosts who have never d/l a successful LHC WU..
Message board moderation

To post messages, you must log in.

AuthorMessage
Travis DJ

Send message
Joined: 29 Sep 04
Posts: 196
Credit: 207,040
RAC: 0
Message 6497 - Posted: 8 Mar 2005, 1:02:46 UTC

Every so often I check the pending credits queue and look at the work units to see how close a given WU is to being complete/granting credit/credit variations or for other more interesting things such as a WU that never gets completed for various reasons, how long did it run, etc.

Occasionally I run across a WU with a lot of download errors and when inspecting the Host's history, they haven't successfully downloaded a single WU since LHC opened back up. For example's sake (and I am in ***no*** way singling anyone out, it was fair chance):

User & Affected WU

Are there any effective ways to notify these users who need to simply download BOINC 4.19 - we do have to give our email addys for registration purposes? It's a waste of LHC's bandwidth if hundreds of megabyes are sent out only to be immediately rejected by the client - but in the same breath, a mechanism to immediately resume WU downloading when the client is "up to date" could help..

Just a suggestion to keep things moving along smoothly :)

ID: 6497 · Report as offensive     Reply Quote
STE\/E

Send message
Joined: 2 Sep 04
Posts: 352
Credit: 1,393,150
RAC: 0
Message 6498 - Posted: 8 Mar 2005, 1:45:01 UTC

I always thought when a WU had a download error it got sent out again right away, obviously not ... The Server should know the WU didn't download and send it out again as soon as possible ... IMO
ID: 6498 · Report as offensive     Reply Quote
Profile The Gas Giant

Send message
Joined: 2 Sep 04
Posts: 309
Credit: 715,258
RAC: 0
Message 6499 - Posted: 8 Mar 2005, 2:18:38 UTC
Last modified: 8 Mar 2005, 2:18:53 UTC

Having a look at the wu in question it has been sent out again almost immediately. Check out the date/time stamps of the error vs the sent of the other wu's. Currently this wu is waiting for 4 hosts to return the wu. I haven't checked the other wu's. But this is one reason why there is a limit on the number of wu's per host per day.

I noticed the error on this wu says signature verification error for both the hosts with the download error and that these hosts are on V4.24. Which is odd since I thought this error was fixed.

Live long and crunch!

Paul
(S@H1 8888)
BOINC/SAH BETA
ID: 6499 · Report as offensive     Reply Quote
Travis DJ

Send message
Joined: 29 Sep 04
Posts: 196
Credit: 207,040
RAC: 0
Message 6502 - Posted: 8 Mar 2005, 5:23:19 UTC - in response to Message 6498.  

> I always thought when a WU had a download error it got sent out again right
> away

The scheduler did send out virtually immediately..

Host..
23655 13:56:47 -> Failed @ 13:57:51
11355 13:58:15 -> Failed @ 13:59:35

With the exception of hosts 21406 & 25836, 5 computers d/l the WU within 5 minutes of each other, 2 failed within 1 mintue of download, then was reissued to the aforementioned two hosts 9 hours later. My best guess in that case is such WUs get "recycled" back into the work queue and go out in order with the rest of the WUs (i.e. they get put back on the top of the stack.. given the low amount of work to go out over the weekend, it *appears* that's what it does).

The Gas Giant is right about the clients in question are v4.24 .. makes me all the more glad I stuck with 4.19 -

However my concern is not about when it failed (timing) and how it was reissued, rather preventing wasted time & bandwidth on the LHC server's part while being able to detect when the condition has been rectified - such as only running sixtrack on stable versions of BOINC.. etc. :)

ID: 6502 · Report as offensive     Reply Quote
STE\/E

Send message
Joined: 2 Sep 04
Posts: 352
Credit: 1,393,150
RAC: 0
Message 6505 - Posted: 8 Mar 2005, 8:27:37 UTC
Last modified: 8 Mar 2005, 8:39:08 UTC

The Gas Giant is right about the clients in question are v4.24 .. makes me all the more glad I stuck with 4.19 -
==========

Yes, there were signature problems at first with any version over v4.19, I even had a bunch of download failures myself. But that didn't stop me from getting WU's, I just switched back to v4.19 until the signature problems were fixed & then switched back to v4.24 & now v4.25 & I haven't had any problems with either one of them ...

In fact the v4.25 can even be downloaded from the Seti Site under the Stable Versions now ...
ID: 6505 · Report as offensive     Reply Quote
Profile PeterHallgarten
Avatar

Send message
Joined: 2 Sep 04
Posts: 14
Credit: 33,774
RAC: 0
Message 6559 - Posted: 14 Mar 2005, 8:07:36 UTC - in response to Message 6505.  
Last modified: 14 Mar 2005, 8:07:54 UTC

> Yes, there were signature problems at first with any version over v4.19, I
> even had a bunch of download failures myself. But that didn't stop me from
> getting WU's, I just switched back to v4.19 until the signature problems were
> fixed & then switched back to v4.24 & now v4.25 & I haven't had
> any problems with either one of them ...
>
> In fact the v4.25 can
> even be downloaded from the Seti Site under the Stable Versions now ...


I had problems with 4.24 with signature errors, a simple reset fixed this issue.

Now running 4.26 now and havent seen any signing issues.
73 de Peter VK3AVE



ID: 6559 · Report as offensive     Reply Quote
Profile Razorirr

Send message
Joined: 18 Sep 04
Posts: 27
Credit: 2,559
RAC: 0
Message 6569 - Posted: 15 Mar 2005, 4:14:39 UTC

i really didnt notice here. were having the same problem at predictor sight now too most of it is me not understanding what their saying but it explains what we are thinking of doing. dont think they would mind you copying. http://predictor.scripps.edu/forum_thread.php?id=1391
sorry your gonna have to cut paste unless someone wants to tell me how to get it to be a hyperlink.
ID: 6569 · Report as offensive     Reply Quote
Profile littleBouncer
Avatar

Send message
Joined: 23 Oct 04
Posts: 358
Credit: 1,439,205
RAC: 0
Message 6574 - Posted: 15 Mar 2005, 11:22:29 UTC - in response to Message 6569.  
Last modified: 15 Mar 2005, 12:20:02 UTC

> i really didnt notice here. were having the same problem at predictor sight
> now too most of it is me not understanding what their saying but it explains
> what we are thinking of doing. dont think they would mind you copying.

THREAD

> sorry your gonna have to cut paste(not in this reply) unless someone(that's me) wants to tell me how to get
> it to be a hyperlink.
>

@ Razorirr, (how to write a hyperlink)
If you open this reply by "Reply to this post": You can see how the hyperlink is written;
Command (HTML-tag): a href= , don't miss brackets and "" !!!
("THREAD" is the example-word for this hyperlink, you can write other discribing words!)

Hope this helps...
greetz littleBouncer

[BTW]on Seti-Forums (also at EAH); when you create a message, you can klick on "You may use HTML tags" (on the left side near your name/avatar); there is a example-list of the tags!


ID: 6574 · Report as offensive     Reply Quote
Ertugrul Gokcen

Send message
Joined: 27 Sep 04
Posts: 22
Credit: 4,026
RAC: 0
Message 6653 - Posted: 21 Mar 2005, 16:15:12 UTC
Last modified: 21 Mar 2005, 16:16:12 UTC

Just take a look at these results!!!

I wonder why that host ever gets WUs!
ID: 6653 · Report as offensive     Reply Quote
Profile Alex

Send message
Joined: 2 Sep 04
Posts: 378
Credit: 10,765
RAC: 0
Message 6663 - Posted: 22 Mar 2005, 6:35:30 UTC - in response to Message 6653.  
Last modified: 22 Mar 2005, 6:36:15 UTC

> Just take a look at these <a> href="http://lhcathome.cern.ch/results.php?hostid=29347">results[/url]!!!
>
> I wonder why that host ever gets WUs!
>

The thing to do is look at a result or two and you'll see

>
4.25
app_version download error: couldn't get input files:

logo_text_right_1.01_.tga
-120
signature verification failed





So, all they have to do is reset, I think.
I'm not the LHC Alex. Just a number cruncher like everyone else here.
ID: 6663 · Report as offensive     Reply Quote
John McLeod VII
Avatar

Send message
Joined: 2 Sep 04
Posts: 165
Credit: 146,925
RAC: 0
Message 6678 - Posted: 23 Mar 2005, 3:08:36 UTC - in response to Message 6663.  

> > Just take a look at these <a>
> href="http://lhcathome.cern.ch/results.php?hostid=29347">results[/url]!!!
> >
> > I wonder why that host ever gets WUs!
> >
>
> The thing to do is look at a result or two and you'll see
>
> >
> 4.25
> app_version download error: couldn't get input files:
>
> logo_text_right_1.01_.tga
> -120
> signature verification failed
>
>
>
>
>
> So, all they have to do is reset, I think.
>
Um, no. What it means is that this file needs to be signed.


BOINC WIKI
ID: 6678 · Report as offensive     Reply Quote
Profile littleBouncer
Avatar

Send message
Joined: 23 Oct 04
Posts: 358
Credit: 1,439,205
RAC: 0
Message 6692 - Posted: 23 Mar 2005, 22:44:00 UTC

One more newcomer on 21. March 2005 with the name: THEFT
THEFT

look here

50 results all errors in one day!
results

greetz littleBouncer


ID: 6692 · Report as offensive     Reply Quote
Profile littleBouncer
Avatar

Send message
Joined: 23 Oct 04
Posts: 358
Credit: 1,439,205
RAC: 0
Message 6731 - Posted: 27 Mar 2005, 18:06:27 UTC

And another one:

over 600 "wrong" results!

Host ID:3220

User: Hatzfelder

greetz littleBouncer
ID: 6731 · Report as offensive     Reply Quote
Astro

Send message
Joined: 17 Sep 04
Posts: 69
Credit: 26,714
RAC: 0
Message 6732 - Posted: 27 Mar 2005, 19:31:42 UTC

Heck, I haven't downloaded a WU of any kind in weeks. lol

tony
ID: 6732 · Report as offensive     Reply Quote
Profile Razorirr

Send message
Joined: 18 Sep 04
Posts: 27
Credit: 2,559
RAC: 0
Message 6733 - Posted: 28 Mar 2005, 3:46:47 UTC

well on this subjuct i ended up detaching from seti because it would give me workunits then i would finish ul and report on time yet for ssome reason it was sayiong that i had 15 bad wus in a month. thats more than they even sent me in that time period!

ID: 6733 · Report as offensive     Reply Quote

Message boards : Number crunching : Hosts who have never d/l a successful LHC WU..


©2024 CERN